1. Fear is the worst advisor.

On March 30, the S&P 500 was down 8% in 2026.
After one of the biggest 4-week rallies ever, it's now up +4.3% YTD.

2. OpenAI is growing but less than expected.

OpenAI is central to the now famous circular AI capex trade.
The WSJ reported that the firm is missing revenue targets and not attracting enough new users.
The news shocked the chip datacenter segments of the economy, who are still spending a ton of cash on the idea that AI tech, and the data centers that run it, will snatch up billions of users and trillions of dollars. The news dropped during yet another pivotal Mag 7 earnings week, five of the biggest names are reporting on Q1 results.
The most important number the market is watching for is CapEx: in the last round of reporting the big boys said they expected to spend upwards of $630B this year on AI.

3. AI models market share.

OpenAI has long been the leader for paid usage by U.S. businesses, but Anthropic has closed the gap with tools like Claude Code and Cowork.
OpenAI and Anthropic have become enterprise software competitors to Microsoft, Salesforce and Servicenow.

4. Buybacks are making a comeback.

"US companies are stepping up share buyback announcements, a sign of balance sheet strength that returns cash to shareholders and helps support stock prices. For investors, that's a constructive signal".

5. Data centers are become bigger and far more expensive.

MS forecasts assume that the cost of rolling out AI data centre capacity will increase from $30mn (per MW, 2023a) to $54mn (by 2030e).
Three years ago, we would have categorised a 1MW data centre as being a large facility. Roll forward to 18mths ago, and that large DC equivalent definition would be 10MW. Today, a 100MW data centre would be labeled as 'large'.

Not a subscriber yet?